|Perl: the Markov chain saw|
Total Perl Dominationby FoxtrotUniform (Prior)
|on Jun 02, 2002 at 21:07 UTC||Need Help??|
So thanks to Ovid (here) and Tilly (here), I've been reading a lot of Paul Graham's Lisp advocacy. In one article, he claims that, among other things, all languages approach Lisp. (I'm not particularly strong on the theory, but IIRC lambda calculus is Turing-equivalent, so he's de dicto correct for any Turing-complete language. But that's not my point.) He talks about how Perl features many Lisp-isms, and how Python is steadily progressing towards apotheosis as Yet Another Lisp Dialect.
Graham's basic, unspoken axiom is that Lisp is the One True Way. Well, we've heard that one before. "Structured Languages are the One True Way." "Object-Oriented Programming is the One True Way." "AI is the One True Way." Yeah, yeah, yeah.
Myself, I suspect that in twenty years, most of the programming will be done by the language, and that most of the time we'll be writing Constraint Logic programs. I'm guessing that our programming environment in a couple of decades will be an optimizing compiler writ large, that can look at a constraint-based description of a problem, determine the best approach, and write (and execute) code to solve it. Sure, a lot of the interesting work will be done at a much lower level, but for most business problems, CLP will be the way to go. (I also think that regexes will be the preferred way to express those constraints.) Who cares what I think? Predicting the future is almost universally a futile exercise.
My copy of High Performance Computing, Second Edition cites an anonymous quote:
"I don't know what language they will be using to program high performance computers 10 years from now, but we do know it will be called FORTRAN."
I can do one better than that:
"I don't know what language we'll be using to solve ubiquitous, practical problems 20 years from know, but I do know it will be called Perl."
Why? Because Perl borrows shamelessly and ruthlessly the idioms and techniques of other languages and of emergent styles of programming, and makes them work together. If CLP really is the way to go, we'll have a Logic::CLP module for Perl 6 and CLP builtins in Perl 7. If a totally different model of computation appears, we'll have that within a couple of releases.
Perl, and its developers, concentrate on making it easy to solve real problems. Stylistic elegance, mathematical brilliance, Byte-headlining buzzwords and, when necessary, programming tradition all take a back seat to that goal.
Perl isn't the One True Way because it's Perl, Perl is as close as we've ever come to the One True Way because it incorporates most of the other, reasonably successful, One True Ways. And instead of sitting back and dogmatically defending its current form, Perl seeks out and assimilates more techniques. Currying. Hyper operators. Anonymous classes (IMHO, the one thing that Java got so very right). Firm typing. More explicit declarations. If it works, we'll take it.
In hoc lingua vinces